espresso machines at 1st-line.com

Grind Sifting for Brewed Coffee - Page 2

Postby shadowfax on Fri Dec 25, 2009 10:18 am

luca wrote:Interestingly, even after sifting, if I brew pourover filter I still seem to get fines on the filter paper. Is it possible that fines are produced by the very act of brewing?

I've noticed a similar thing when cupping and filtering the coffee through glass rod/metal mesh filter, and my guess is that these particles are so small and so statically charged that they don't separate during sifting. Of course, water would neutralize that static electricity... Anyway that was my thought here:
shadowfax wrote:These brews still let some silt through, oddly enough, but it's very strange: it's so ultra-fine on my fingers that I can barely feel it, and drinking the silty bottom, while unpleasant, is a lot more palatable than drinking the silt from a brew of unsifted grinds. I am thinking this silt is from extremely, extremely tiny particles that are statically attached to the big particles, where the static charge is stronger than the force of the agitation of the sifter, given their tiny mass.
Nicholas Lundgaard
User avatar
shadowfax
Team HB
 
Posts: 3301
Joined: May 04, 2005
Location: Houston, TX

Postby another_jim on Sat Dec 26, 2009 10:46 pm

shadowfax wrote:Has anybody out there messed with sifting/sieving the fines from dry coffee grinds before brewing? ... I've always been rather bugged by the silt in my French press coffee ...


Just a slight off topic note. Removing fines will help with the silt problem. However, it should have no effect on the taste.

Fines are broken pieces of coffee cell wall, while coarse particles are one or more complete cells. The cell walls are insoluble cellulose, and brewing them results in no taste whatsoever. When the cell wall is broken, the soluble and volatile contents of the cell get smeared over the rest of the ground coffee, and this contributes to over extraction. I cannot see how sieving out the cell walls, while leaving the solubles, will change this.

In other words, you cannot get the same taste as a grinder that produces fewer fines by sieving one that does produce a lot of fines*. However, given the particle distributions of decent grinders, not that many cells are destroyed, and the difference may not be very distinguishable.

* Unless the bulk of the released solubles is sticking to the fines, but not the coarse particles, which would be odd.
User avatar
another_jim
Team HB
 
Posts: 9142
Joined: May 05, 2005
Location: Chicago
kshanticoffee.com: we love to make things and we also love coffee
kshanticoffee.com: we love to make things and we also love coffee

Postby shadowfax on Sun Dec 27, 2009 12:04 am

Interesting; How would you go about testing this well, Jim? If your theory is correct, am I to assume that I should be pre-weighing my coffee doses and then sifting samples and comparing sifted vs. unsifted, where the sifted samples weighed the same as the unsifted ones prior to sifting? My point being, if fines don't contribute to the extraction, I wouldn't be doing an accurate comparison if I compared a sample with the fines removed and the dose 'increased' such that the sifted sample had equivalent weight to the unsifted one. This is what I did in the original non-blind comparison I did with Black Cat, which may have contributed to the obvious difference we tasted in the cup.

On the other hand, as I've continued to use the MPS-50, I get the strong impression that it does indeed remove more than just the fines, and seems to remove also the smaller non-fine particles, which explains why there is such heavy particle loss at fine-ish grinds. If there's a correlation that the fines are broken off mainly from the smaller particles (i.e. mostly the smaller non-fine particles have broken cell walls), then it could be that such sifting would improve the cup. We sure noticed a balance shift in the coffees we sampled, but this could have, I suppose, been the mark of higher dose + underextraction making the cup a little brighter.

I suppose that muddy 'flavor' is more of a texture thing after all...
Nicholas Lundgaard
User avatar
shadowfax
Team HB
 
Posts: 3301
Joined: May 04, 2005
Location: Houston, TX

Postby another_jim on Sun Dec 27, 2009 12:39 am

For comparing just taste, I would recommend

-- starting out with the same weight of ground coffee and water prior to any sifting.
-- sift one sample, and leave one or two unsifted
-- steep them all, then paper filter them.

This will give you (close to) a taste comparison rather than a mouthfeel + taste one. The (close to) is because the sample with fines will drain through the paper filter more slowly. If you can do a triangle test, using two unsifted cups, one sifted one, and pick out the sifted one, it will prevent a false positive based on slight cup to cup variation (which I always seem to get).

The other way, which I now prefer to triangle testing is to taste the two alternatives for a while, either blind or not, and see if you can learn the difference, and then pick them out blind consistently.
User avatar
another_jim
Team HB
 
Posts: 9142
Joined: May 05, 2005
Location: Chicago

Postby hbuchtel on Mon Dec 28, 2009 2:52 am

I've brewed coffee using the 'fines'* sieved out using the MPS, both as espresso and cold brew, and there is plenty of flavor to be extracted from the grinds.

This does not contradict Jim's "brewing [cell walls] results in no taste whatsoever" idea, as the MPS's mesh size... is kinda big.

How about Turkish coffee? The grind is basically powder, but the coffee certainly has flavor...

Regards, Henry

*the current definition of 'fines' seems to be "the grounds that are much smaller than the size that I wanted", which is not particularly precise... I like the idea of defining them as "non-intact cells", but it will be impossible for the average person to identify fines vs very fine grounds.
LMWDP #53
User avatar
hbuchtel
 
Posts: 752
Joined: Jun 22, 2005
Location: Ann Arbor, MI

Postby another_jim on Mon Dec 28, 2009 4:34 am

The fines peak on the distribution diagrams consists of broken cells. But the coarse particle distribution covers a 10* size difference. The scale on the distribution curves is logarithmic, which masks the size differences. Unfortunately, a physical screen set cannot copy the neat trick of going in logarithmic steps, so the sieved sizes will cut the big distribution.

Does a particle with 10 complete cells brew any slower than one consisting of just one complete cell? Does grind size, providing it's fine enough that all the cells have good wetting, matter much for steeping methods?

When I tested the big Ditting (see curve at the top of his topic) against various espresso grinders at Terroir, the Ditting spanked them. But all the espresso capable grinders I've compared at home (the big Ditting doesn't produce enough fines for espresso) have not shown clear differences, not even going from big commercial conicals to the small home grinders.

I'm not making statements here, just asking questions. Quite honestly, I have no real grasp of what creates a better grind for brewed coffee. I think the simple idea of all the particles having to be roughly same size may not stand up to close examination. Particles shaped and packed so that all the coffee cells wet equally well may be a better way of thinking about grind quality.
User avatar
another_jim
Team HB
 
Posts: 9142
Joined: May 05, 2005
Location: Chicago

Postby RapidCoffee on Mon Dec 28, 2009 11:20 am

hbuchtel wrote:I've brewed coffee using the 'fines'* sieved out using the MPS, both as espresso and cold brew, and there is plenty of flavor to be extracted from the grinds.

This does not contradict Jim's "brewing [cell walls] results in no taste whatsoever" idea, as the MPS's mesh size... is kinda big.

This makes sense to me. I don't believe there is any evidence that coffee "cells" (in either the biological or structural sense) survive the roasting and grinding process intact. In fact, SEM pix suggest the opposite. However, if the MPS is sieving out ~20% of the coffee by weight, the mesh size must be significantly larger than the ~40um diameter of the fines peak.

The most compelling argument that fines have little direct contribution to taste comes from sheer numbers. Particle sizing results indicate that fines comprise just 1-2% (at most) of the coffee mass. (Either that, or I'm interpreting the plots incorrectly. :? )
John
User avatar
RapidCoffee
Team HB
 
Posts: 3169
Joined: Dec 11, 2005
Location: Rapid City, SD
Quotable Quote
"It's something of a zero-sum game -- you can learn to pull an occasional good shot on very inexpensive equipment, however, it will take a long, long, long time. On the other end of the scale, I could teach you to pull a decent shot on a La Marzocco and Super Jolly in just a few minutes." --Jeff, Is home espresso better?


Postby portamento on Mon Dec 28, 2009 1:15 pm

You can estimate the mesh size by counting the number of openings across one linear inch. Wire thickness plays a role too so this is not a precise method.

Let's say you count 30 openings. That's 30 mesh which according to a conversion chart corresponds to 595 microns.
portamento
 
Posts: 316
Joined: Jan 25, 2008
Location: Texas

Postby cannonfodder on Tue Dec 29, 2009 12:07 am

Sizing can be complicated. To many standards. There is grit, mesh, new Japanese standard, old Japanese standard, etc... The only way to really level the playing field is to use micron. 8000 grit is heavier than 8000 mesh which is heavier than 8000 jpn but a micron is a micron the world around.
Dave Stephens
User avatar
cannonfodder
Team HB
 
Posts: 7977
Joined: May 23, 2005
Location: Downingtown PA

Postby CRCasey on Tue Dec 29, 2009 12:32 am

But when you have to go to a fractional wavelength of a absorption pattern of light don't you think that a size that small really is not that useful in fluids like this. That is not what we want to measure for. You can pick a standard and work with it. A Micron standard may be overkill for this.

-Cecil
Black as the devil, hot as hell, pure as an angel, sweet as love-CMT:LMWDP#244
User avatar
CRCasey
 
Posts: 675
Joined: Jan 20, 2009
Location: Lewisville, TX